The July Effect

Every July, a smattering of news outlets remind the public that July 1 is when newly minted doctors, fresh out of medical school start out at teaching hospitals across the country. Most commonly, the articles tell everyone to avoid the hospital because clearly, having new doctors around is dangerous and deadly. This year, I came across Lifehacker highlighting the 'July effect' in this post.

As you can see from the comments, there is no shortage of people with horror stories about medical encounters. I certainly understand such frustration as I have also experienced such events with myself and my family. Many of these events, however, may have nothing to do with what really happens during the July transition. For example, one person talks about a bad attempt at having an IV started, but doctors generally do not start IV's (I can count on one hand how many times I have started an IV). Some others point out that they have had excellent experiences being treated by interns and medical students. As a former intern, I remember my first procedures and I would argue that my trepidation might have made me more attentive than the seasoned doctor who, after doing hundreds of procedures, is complacent about some details. For that matter, in an appropriate teaching setting, anyone doing their first few procedures is directly observed by another doctor who is significantly more experienced. Admittedly this is not always the case, but it is how the system is supposed to work.

Okay, so now that we have talked about some of the circumstantial evidence, is there any actual evidence of a 'July Effect'? One of most oft cited studies was released a year ago in the Journal of General Internal Medicine, which found an increase in fatal medication errors in July. The authors reviewed death certificates and found a 10% increase in risk consistently during July and was seen only in counties with teaching hospitals. The really unfortunate thing is that the manuscript only reports the RELATIVE risk increase, not the ABSOLUTE risk increase. For example, we do not know if fatal medication errors increased from 1% to 1.1% or was the increase from 0.00001% to 0.000011%. Both of these are 10% relative increases but the former is obviously much more meaningful than the latter. This study is also limited by including data back to 1979 when teaching hospitals and supervision of residents was quite different than modern teaching systems.

This study however, is one of the only ones that actually found evidence of an effect. Many more studies using a variety of different data sets found no evidence and in fact, Wikipedia actually has a pretty comprehensive list of the studies examining this phenomenon. Somehow, though, these negative studies do not seem to get the overwhelming media attention that the one positive study did (Google search for July effect).

Of course, the other problem is that while sometimes you can pick when to have your knee replaced or to have a face lift, I do not know anyone who chooses when they have a heart attack or get pneumonia. Furthermore, there is evidence that other times are equally, or more concerning. This analysis by Cavallazzi et al from CHEST in 2010 showed that ICU admissions at night were not a greater risk, but being admitted over the weekend had a 8% increased relative risk of death.

Is there a July effect? If you are at a teaching hospital, you are likely to encounter at least one doctor fresh and shiny out of medical school, but there will be experienced nurses, higher level trainees, and supervisory doctors checking up on the newbies and making the important decisions which are left to the intern to carry out. Does this make you more likely to die as a result of being cared for by an intern? I seriously doubt it.